Chapter 5 202 6. Power considerations for validation of the test on validation set 1, validation set 2 and the chemotherapy cohort For each sample set, the power to detect the effect sizes between test classification groups observed in the development cohort was estimated using the proportions of each classification observed in development and the proportions of each classification observed in the sample set upon testing. All calculations assume fully mature clinical data and α=0.05, two-sided. i. Validation set 1 (n=98) The expected number of patients of each classification group given development set data are 35 resistant, 36 intermediate and 27 sensitive. On running the test on the samples of validation set 1, 37 samples were classified as resistant, 30 as intermediate and 31 as sensitive. Supplementary Table S16: Power to detect effect size observed in the development set in validation set 1. PFS OS Comparison Classifications predicted Classifications assigned Classifications predicted Classifications assigned Resistant vs. intermediate 60% 57% 49% 47% Intermediate vs. sensitive 57% 57% 62% 62% Resistant vs. sensitive 99% >99% 99% 99% Resistant vs. not resistant 96% 96% 94% 94% Sensitive vs. not sensitive 94% 96% 94% 96% ii. Validation Set 2 (n=75) The expected number of patients of each classification group given the development set data are 26 resistant, 28 intermediate and 21 sensitive. On running the test on the samples of validation set 2, 32 samples were classified as resistant, 19 as intermediate and 24 as sensitive.
RkJQdWJsaXNoZXIy MTk4NDMw